Astron. Astrophys. 345, 244-248 (1999)

3. Abundance analysis

Given the instrumentation set-up, the abundance analysis from high-resolution spectra is based on the range 4235-4270 Å while low resolution spectra cover the 3952-4938 Å interval. Thus the optical region is fairly well covered; however, photospheric abundances are in general difficult to determine due to the weakness of the majority of photospheric lines. The UV resonance lines are much more pronounced but require exhaustive NLTE treatment. To start with, we derived carbon abundance and values from the 4268 Å line profile. The subsequent analysis is based on the best atmosphere models for early type stars, i.e. Kurucz (1992) models.
It turned out that it was better to put aside the constants for Stark broadening for the C II 4268 line and to apply instead the classical formula, , where stands for electron plasma density and the effective quantum number of the upper level; otherwise, the line wings did not fit any of the observed line profiles. This effect was remarkably well established thanks to several cases where the value is equal to zero. Other damping constants (e.g. natural radiative) associated with broadening mechanisms matched the observations very well. Other line parameters, oscillator strengths etc., have been maintained from the original Kurucz-Peytremann line list. Yet, tests have been made using other sources for atomic parameters, with values increased or decreased by 50 percent, but the pattern of CNO abundances did not change. The abundance analysis has been performed by a standard synthetic spectrum and LTE method. The number of well pronounced line profiles for light elements in the spectrum is not large enough to estimate reliably the microturbulent velocity and therefore we adopted a depth-independent microturbulent velocity value of 6 km s^-1 (Kilian 1992). The abundance analysis is based on several lines reaching only a few percent of the continuum level (Fig. 1) which, unfortunately, were found to depend on basic stellar parameters, i.e. effective temperature and surface gravity (Fig. 2). This demonstrates already how fragile is the construction of CNO abundances vs. stellar age diagram. The analysis was primarily focused on prominent CNO photospheric lines while weaker as well as other lines (e.g. S, iron peak) will be analysed later except for the analysis of the prominent Mg II 4481 line profile.

Fig. 1. A portion of the high-resolution spectrum of the reference star HD 122980 (dots ) compared with synthetic spectra. Solid line: fits of N, O, Mg, S abundances, dashed line: solar abundances,  km s-1. See text.

Fig. 2. A portion of synthetic spectrum vs. basic stellar parameters. Solid line: T K, , dashed line: 20000 K, 4.0 and dotted line: 19000 K, 4.5.

© European Southern Observatory (ESO) 1999

Online publication: April 12, 1999
helpdesk.link@springer.de